IoU-balanced Loss Functions for Single-stage Object Detection

Wu, Shuang; Yang, Ji; Wang, Xinggang; Li, Xiaoping

doi:10.48550/arxiv.1908.05641

Cited by 17 publications

(24 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As an excellent anchor-free detector, FCOS [20] predicts centerness scores to suppress the lowquality detections. Instead of adding an extra branch, IoUbalanced classification loss [24] and PISA [3] adopt respectively the normalized IoU and sorted score to reweight classification loss based on their localization accuracy, which strengthens the correlation of classification confidence and localization quality. Besides, several works aim to design a joint representation of localization accuracy and classification.…”

Section: Related Workmentioning

confidence: 99%

“…Learning to predict offset. To bridge the gap from classification score to localization confidence, we adopt IoU between predicted bounding boxes and corresponding ground-truth boxes as localization confidence as the same as [24,10]. Fortunately, the criterion(IoU) is easily calculated by the 4 predicted offsets between predicted boxes and corresponding anchors.…”

Section: Offset Prediction Between Classification and Localizationmentioning

confidence: 99%

See 1 more Smart Citation

SCD: A Stacked Carton Dataset for Detection and Segmentation

Yang¹,

Wu²,

Gou³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Carton detection is an important technique in the automatic logistics system and can be applied to many applications such as the stacking and unstacking of cartons, the unloading of cartons in the containers. However, there is no public large-scale carton dataset for the research community to train and evaluate the carton detection models up to now, which hinders the development of carton detection. In this paper, we present a large-scale carton dataset named Stacked Carton Dataset(SCD) with the goal of advancing the state-of-the-art in carton detection. Images are collected from the internet and several warehourses, and objects are labeled using per-instance segmentation for precise localization. There are totally 250,000 instance masks from 16,136 images. In addition, we design a carton detector based on RetinaNet by embedding Boundary Guided Supervision module(BGS) and Offset Prediction between Classification and Localization module(OPCL). OPCL alleviates the imbalance problem between classification and localization quality which boosts AP by 3.1% ∼ 4.7% on SCD while BGS guides the detector to pay more attention to boundary information of cartons and decouple repeated carton textures. To demonstrate the generalization of OPCL to other datasets, we conduct extensive experiments on MS COCO and PASCAL VOC. The improvements of AP on MS COCO and PASCAL VOC are 1.8% ∼ 2.2% and 3.4% ∼ 4.3% respectively.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Offset Prediction Between Classification and Localizationmentioning

confidence: 99%

SCD: A Stacked Carton Dataset for Detection and Segmentation

Yang¹,

Wu²,

Gou³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Can we bias the model more to the features aligning better with the objects? Inspired by IoU-balanced localization loss [22], we design GGIoU-balanced localization loss to realize this goal.…”

Section: (B))mentioning

confidence: 99%

Gaussian Guided IoU: A Better Metric for Balanced Learning on Object Detection

Wu¹,

Yang²,

Yu³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

For most of the anchor-based detectors, Intersection over Union(IoU) is widely utilized to assign targets for the anchors during training. However, IoU pays insufficient attention to the closeness of the anchor's center to the truth box's center. This results in two problems: (1) only one anchor is assigned to most of the slender objects which leads to insufficient supervision information for the slender objects during training and the performance on the slender objects is hurt; (2) IoU can not accurately represent the alignment degree between the receptive field of the feature at the anchor's center and the object. Thus during training, some features whose receptive field aligns better with objects are missing while some features whose receptive field aligns worse with objects are adopted. This hurts the localization accuracy of models. To solve these problems, we firstly design Gaussian Guided IoU(GGIoU) which focuses more attention on the closeness of the anchor's center to the truth box's center. Then we propose GGIoUbalanced learning method including GGIoU-guided assignment strategy and GGIoU-balanced localization loss. The method can assign multiple anchors for each slender object and bias the training process to the features well-aligned with objects. Extensive experiments on the popular benchmarks such as PASCAL VOC and MS COCO demonstrate GGIoU-balanced learning can solve the above problems and substantially improve the performance of the object detection model, especially in the localization accuracy.

show abstract

“…In the past few years, with the development of Convolutional Neural Networks, object detection has been applied in many scenes, such as security, autonomous driving, defect detection. The CNN-based detection models are mainly classified into single-stage detectors [6,14,5,4,15] and two-stage detectors [3,16,17]. Faster R-CNN [3], acted as the mainstream two-stage detector, firstly uses RPN to generate region proposals which are fed into the FPN to conduct accurate localization and classification tasks.…”

Section: Object Detectionmentioning

confidence: 99%

Carton dataset synthesis method for domain shift based on foreground texture decoupling and replacement

Gou¹,

Wu²,

Yang³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

One major impediment in rapidly deploying object detection models for industrial applications is the lack of large annotated datasets. We currently have presented the Sacked Carton Dataset(SCD) that contains carton images from three scenarios such as comprehensive pharmaceutical logistics company(CPLC), e-commerce logistics company(ECLC), fruit market(FM). However, due to domain shift, the model trained with carton datasets from one of the three scenarios in SCD has poor generalization ability when applied to the rest scenarios. To solve this problem, a novel image synthesis method is proposed to replace the foreground texture of the source datasets with the foreground instance texture of the target datasets. This method can greatly augment the target datasets and improve the model's performance. We firstly propose a surfaces segmentation algorithm to identify the different surfaces of the carton instance. Secondly, a contour reconstruction algorithm is proposed to solve the problem of occlusion, truncation, and incomplete contour of carton instances. Finally, the Gaussian fusion algorithm is used to fuse the background from the source datasets with the foreground from the target datasets. The novel image synthesis method can largely boost AP by at least 4.3% ∼ 6.5% on RetinaNet and 3.4% ∼ 6.8% on Faster R-CNN for the target domain. And on the source domain, the performance AP can be improved by 1.7% ∼ 2% on RetinaNet and 0.9% ∼ 1.5% on Faster R-CNN. Code is available here.

show abstract

IoU-balanced Loss Functions for Single-stage Object Detection

Cited by 17 publications

References 0 publications

SCD: A Stacked Carton Dataset for Detection and Segmentation

SCD: A Stacked Carton Dataset for Detection and Segmentation

Gaussian Guided IoU: A Better Metric for Balanced Learning on Object Detection

Carton dataset synthesis method for domain shift based on foreground texture decoupling and replacement

Contact Info

Product

Resources

About